We build for markets too small to matter before: artisanal mining, small-scale farming, in-country research, local civil society, community interventions, remote value chains. Banking the unbanked. Small-scale climate projects, small-scale conservation, small-scale reforestation. The list keeps growing, and we will come back to why that is the point rather than a problem.
Notice what the list does not contain. Not one country, not one region, not one continent. That is deliberate, and it is the actual thesis of this post: what makes a market small enough to be overlooked is a shape of problem, not a place.
What "too small to matter" actually meant
None of these markets was ever small in the lives it holds. Artisanal mining and small-scale farming employ enormous numbers of people. Local civil society does work that no one else will do. "Too small to matter" was never a judgment about human importance. It was a judgment made by the economics of traditional software.
Traditional data infrastructure is built for buyers with procurement budgets, standardised processes, and enough seats to amortise an enterprise licence. Where those conditions hold, vendors compete fiercely. Where they do not, nothing gets built at all. The result is a familiar picture: the knowledge exists, but it sits in a person's head, on paper, or in a spreadsheet nobody else can read. Sometimes it is structured but siloed, undiscoverable to the very people who would act on it. The information is real. The infrastructure to carry it never arrived.
That is our market. We define it by three conditions, not by a map: digital infrastructure that is really backward, real room to improve the way digitisation happens, and information that is scarce and hard to get. Wherever those three conditions hold together, we are interested. Wherever they do not, we are not, however familiar the geography might look.
Why the axis is sector, not geography
It would be easy, and it is common, to name this market by pointing at a region. We refuse to, for two reasons.
The first is accuracy. The three conditions recur on every continent. A remote value chain behaves like a remote value chain whether the remoteness is a rainforest, a mountain range, or simply distance from the nearest institution that keeps records. Community health worker networks, water user associations, informal cross-border trade, community savings groups: the shape repeats wherever formal infrastructure thinned out before it reached the people doing the work. Drawing a border around the market would exclude places that fit and include places that do not.
The second is respect. Naming a geography quietly turns a market into a charity case, a place to be developed rather than a set of people already doing sophisticated work without the tools they deserve. Naming a sector does the opposite: it says the work itself is the market. An artisanal miner, a smallholder farmer, and an in-country researcher have more in common with each other, as users of information infrastructure, than any of them has with a neighbour who happens to work inside a formal institution across the road.
So when the list of markets grows, it grows sector-wise. Each new entry is another instance of the same three conditions, never another pin on a map.
The blank sheet
Here is the strange advantage of building where infrastructure never arrived: there is nothing to displace. We do not compete with traditional data infrastructure in these markets because it simply does not exist. It is a blank sheet of paper. That changes what building means. There is no legacy system to integrate with, no incumbent to unseat, and, more importantly, no inherited assumptions about who data belongs to.
On a blank sheet, you get to write the ownership rules first, and we write them one way: the people who hold the knowledge stay custodians of it. What we build unlocks data that either never existed in structured form or existed but sat siloed, renders it into one shared, comparable form, and opens a protocol through which the people who hold it reach the people who need it. We broker that encounter, permissioned and direct, and we take a margin on the exchange we facilitate. Value realised, not extracted, is the constraint on every part of that sentence. Participants take part at their own discretion, to their own benefit. Their data is not extracted but elevated. They do the work and own the outcome; we are building infrastructure that promotes them.
That framing is not decoration. In markets where the first arriving infrastructure sets the terms for everything after it, whoever writes on the blank sheet first decides whether data flows away from the people who produced it or works for them. We think the second version is both the right thing and the better business, because infrastructure people trust is infrastructure people actually use.
An honest hedge
We will be candid about one thing: extending our positioning sector-wise rather than geography-wise is a strong leaning, not a finished doctrine. We are still testing where the list of markets ends, and whether a growing list can carry the positioning without collapsing back into a map in the reader's head. Some of these sectors we know deeply from years of work; others we recognise by their shape and have yet to build for. That uncertainty is material, so it belongs in the post.
But the core claim we are sure of. The markets we build for were never too small. They were too small for a particular economics of software, and that economics was never a law of nature. Three conditions, not a map. Wherever they hold, the sheet is blank, and the people who hold the knowledge should be the ones who benefit when it finally gets written down.